Provides introductions and tutorials on Bilibili-related technologies.
This article details Bilibili's third‑generation streaming transcoding system, explaining its architecture, core modules, testing strategies, quality‑assurance plans, and code examples that together dramatically reduce processing time and support millions of daily video uploads.
This article explains the technical challenges and innovative AI solutions behind preserving speaker identity, emotion, and timing while translating video content into multiple languages, covering speech generation modeling, speaker segmentation, adversarial reinforcement learning, proper‑noun adaptation, and audio‑visual alignment techniques.
This article details Bilibili's multi-layer storage architecture and Alluxio‑based cache acceleration for large‑scale AI model training, covering challenges of high‑throughput, low‑latency file access, metadata scalability, fault tolerance, and the engineering solutions that boosted I/O performance up to ten‑fold.
This article introduces a language‑centric tree reasoning (LTR) framework that recursively decomposes VideoQA queries into perceptual sub‑questions and performs bottom‑up logical inference with video assistance, achieving significantly higher accuracy and explainability across eleven benchmark datasets.
IndexTTS2 introduces a cross‑modal, multi‑language voice translation system that preserves speaker identity, acoustic space, and multi‑source timbre, while tackling challenges like voice personality loss, subtitle cognitive load, localization costs, multi‑speaker diarization, and cultural adaptation through novel time‑coding, adversarial RL, and diffusion‑based lip‑sync techniques.
This article details the comprehensive redesign of a mobile video‑editing page, covering background challenges, requirement analysis, MVVM + UDF architectural choices, module decomposition, dependency‑injection implementation, and the design of a robust Redo/Undo system to improve maintainability and user experience.
This article introduces the comprehensive design and evolution of a unified metadata lineage platform for big data, covering background, data processing chain, lineage models, system architecture, quality metrics, application scenarios, and future plans to enhance data governance, quality, and security.
This article describes the design and implementation of a new frontend material platform that consolidates npm packages, UMD scripts, and icons, introduces a custom publishing tool (bpm), integrates with Nexus3, adds unified authentication, improves search with MeiliSearch, and outlines migration, classification, and documentation strategies.
IndexTTS2 introduces a novel auto-regressive zero-shot text-to-speech model that achieves precise duration control and fine-grained emotional expression through a universal time‑encoding mechanism, decoupled voice‑style and emotion modeling, and a GPT‑style latent feature, outperforming state‑of‑the‑art baselines across multiple benchmarks.
This article explains why heterogeneous servers cause CPU performance layering, describes how to detect the issue using metrics such as NUMA hit/miss rates, cache miss ratios and frequency states, and provides step‑by‑step remediation techniques—including NUMA binding, cache isolation, recompilation and frequency locking—to improve resource pooling efficiency in modern data centers.
